GO BACK to INDEXDYNO ACCURACY, Comments:

Almost Anything that can absorb energy from an engine (or Prime Mover) can be used as a Dynamometer. A vehicle itself can be used as a Dynamometer simply by measuring the time needed to increase its speed (against mass and drag). The vehicle's Mass absorbs the engine power by increasing its Kinetic Energy (Speed). This is considered a Kinetic Dyno, just like any other inertia type Dynos (engine or Chassis). A Static Dyno is one which can continuously Absorb engine power while able to hold the engine at a Set Static Speed. Usually, the energy absorbed is eventually turned into Heat energy. Water has the ability to absorb heat very well and is widely used with dynamometers. Electric Dynos may use magnetic fields to absorb and convert energy into heat, but they usually need water for cooling. Each type of dyno (static and kinetic) has its merits and the 'Ideal Dyno' might be one that can be both.

A Dynamometer Abosorbs POWER (energy per time). There must be a perfect balance between power/energy created by the engine's combustion process and everthing that absorbs the power/energy (Law of Physics). There is no law that states that the dyno and engine torque must be equal.

Water Brake Dynos: 'Common sense says' that the engine torque output of the engine MUST equal the Dyno Arm Torque (water brake dyno at steady state). Not true, and is only a coincidence if it does seem to be equal. All we can ever expect is that the Dyno Arm torque 'errors' are consistent from pull-to-pull so that we can detect changes in engine performance. The Dyno Absorbs Energy (or Power) and there Must be an equal balance of power generated and absorbed (law of conservation of energy). Some dyno designs are more consistent in absorbing power. Any water brake will have significant errors when the internal water temperature is high and flashes to steam internally and escapes. The Steam is now absorbing Energy that does not show up on the dyno torque arm through reduced momentum transfer. You will typically find water brake dynos that produce different results from the same repeatable engine, even though each dyno is very repeatable.

Dynamo and Eddy Current Dynos: Provide a very predictable 'transfer' of Torque from input shaft to output dyno arm, although they typically have high inertia and drive shafting can be very tricky. High inertia results in large torque arm errors if the engine/dyno speed is not steady.

GENERAL:
Measuring the RAW (Brake) Power generated by an engine can be done very Consistently, or more repeatable than most testing standards require. (The DEPAC system has shown changes as small as 0.001%). Power is the rate of change of Energy, and this can be measured directly or indirectly. Engine Torque can be determined from this measurement of Power.

The common Water Brake can have errors if you allow the internal water temperature to become too High. When Steam is created inside, and Exits the dyno, the Torque measured will drop and become erratic. Bad placement of Water In/Out hoses will also cause errors. Test for this by simply turning the water ON , with no engine, to see these 'Flow' caused errors. (Why?) Even in expensive electric dynos, the internal cooling water flow can generate false torques (documented on GE eddy current dynos). It is important to understand the characteristics and limitations of your dyno.

Most all Dynamometer brakes measure Torque off the 'free to pivot' absorber. This is an IN-DIRECT method for measuring engine torque load. It has always been assumed that under steady RPM loads that this Torque measurement was Equal and Opposite the Engine's torque output. This is NOT VALID! Consider that the connecting driveshaft can absorb power through windage and whipping. These losses will reduce the torque measured at the brake. Also different internal designs for water brakes will cause differences in torque readings on the same engine. The Dyno/Brake Absorbs Power not just Torque. Thus you cannot compare results between different style brakes (several %), although you may be able to see 0.01% changes on any one brake.

Although we have shown that engine/dyno measurements made on one dyno setup can be very repeatable, the Absolute accuracy can be very poor. Its not uncommon to see the same engine make different numbers on different dynos, although very repeatable on any one.

Measuring Engine Torque Load off the Dyno brake is Traditionally accepted, BUT is Not Accurate!

It is BEST to measure the Engine Torque DIRECTLY off the Drive shaft (with inertia corrections applied). This method can be Absolutely Accurate if applied properly! It is about time that dyno testing standards advance to this century.



Other Considerations for Engine/Dyno Accuracy: CORRECTION FACTORS

Consistent Testing procedures are Needed so that different engine tests can be compared. A system of Consistent Testing Procedures and Correction Factors form the Basis of Most Engine Testing Standards.

Trying to correct for different Atmospheric conditions using Correction Factors (CF) is the next Most In-Accurate part of engine testing. Understanding that Correction factors, applied to the measured Engine Torque and Power, are far from perfect is very important. Certainly, Do Not Blindly Depend on them. (See, a more detailed discussion of Atmospheric Correction Standards, Link below).

Correction Factors were originally determined by experiment. Different engine types were found to need different CF formulas. You might think at first that a CF number would simply be related to Air Density, but this is not true. The CF formula Set by each Standards Committee (SAE, ISO,,) was just a Compromise Approximation to specify one 'Magical' Formula for all CFs. This Approximation, though, is acceptable if the CF standard is used properly. If you are correcting for small changes in the atmospheric conditions, the CF can be very meaningful. Avoid using the CF formula to correct results for large changes in the Atmospheric conditions from the Standard.

Consider the case of two conditions (A&B) that have Equal CF but where condition (A) has a Low Barometer and a Low Air temperature and condition (B) has a High Barometer and a High Air temperature. The Same CF Implies that the engine should produce the Same Corrected performance. Experience proves that the engine performs differently under conditions A&B..... Get the Hint?.. Control your conditions if you want or need consistent test results!

Of the 3 factors in the CF equation AIR Temperature is the Most Important. It has a Major effect on combustion and fuel vaporization quality. You can Not get an Alchy engine to run at freezing air temperatures. (Soo, what meaning has the CF now??). You should try to CONTROL the Air Temperature as close as possible and fortunately this is not that difficult. Moisture in the air is the next important fuzzy variable. Lotsa moisture can affect mixture quality/ratio and can have a quenching effect (you can use more spark advance with 'heavy' moist air). Its important to Know the effect of air moisture on engine tuning at the track/event but you should try to test with a minimum amount of air moisture to reduce its variability on results. Taking water out of the air is not that difficult and in the winter months the air is also dry to start. The Barometric pressure is the most consistent part of the CF equation and its the only factor we should correct for while we hold the air temperature and moisture constant.

You can Accurately measure engine performance in one set of conditions but you can not accurately convert them to another very different Atmospheric condition. But you can have good Relative Accuracy when comparing results of small changes. If you find that a change causes a 1% (relative) Increase in engine torque at a given speed, then you can expect a very similar increase under a different set of conditions. Most of all Engine testing is looking for Differences caused by changes that affect performance, efficiency, or emissions. So, although it may be difficult to translate results, Absolutely, from one set of conditions to another, you can be assured that small changes that improve a result in one set of conditions will translate well into another set of conditions.

The real problem with the use of Correction Factors is that someone will always mis-use them. Someone builds an engine that's peaked to run best at sea level pressure. Now take that engine to Mexico City at 8,000 feet. Will it now run at its best and can you depend on the correction factor to predict performance?. At 8,000 feet density altitude you can increase the engine's Compression ratio to more than 19/1 for best performance/efficiency. How does the CF apply in these extreme conditions?

We propose changing the Correction Standards that allow testing under local conditions, as a Local Standard, and then specifying these conditions as an Required part of the results (like at 8,000 feet). If someone wants to translate the test results to another set of conditions, then they are free to do so, as long as the original results are left intact. For example, there could be three columns in the dyno results for Uncorrected, Corrected to the local standard, and Corrected to another, more universal, standard like SAE J1349. This last column would attempt to standardize the results to a common standard but this would be the least accurate of the three results.

If your test conditions are fairly close to an established Atmospheric Standard, say within +/- 2.5%, then you could probably adopt this standard, say J1985 with 29.53"Hg with 1% moisture pressure at 25C (77 DegF). But if you are at 6,000 Ft altitude then correcting to your Mean Local weather conditions would be best.

For Best Accuracy, It would be better to Control the Atmospheric Conditions of the Engine Test rather than to use Correction Factors. CONTROL of ALL test conditions is Key to the most Accurate (or Meaningful) Dyno Results.

(See, a more detailed discussion of Atmospheric Correction Standards).

BACK to INDEX

DYNO TESTING CONDITIONS:

For Meaningful Dyno Results, you should test engines as close possible to the intended application. This should be very obvious but so often this common sense rule is ignored. Why even bother to test an engine if you cannot simulate the application (and all of its potential problems). You may get the engine to Scream on the dyno only to Stumble in the race car.

Even if the dyno is used just to break-in an engine, the setup is still important. A Good Example.. Many people use a coolant (tower) system that mixes cool tap water with the warm engine water using a thermostat. A common and simple set-up, But Danger! You may be breaking-in a rebuild that has a compression leak in the head gasket. This simple cooling system just lets these bubbles exit through the top, open to the air. You are oblivious to this serious problem. You put this engine in the car and promptly blow out the radiator water and possibly fry the head(s).

If you used a cooling system that was closed, and used the same type of Pressure cap that's used in the race car, you would have immediately seen this problem 'on the dyno' and not at the track. We would like to see heat exchanges being used all of the time, using either water or air to carry the heat away. Even an old Brass Radiator would work fine, and even better if you could also simulate the same plumbing as in the car to check for pump cavitation or other cooling problems.

It should be very obvious that you should also use the same (type) of Exhaust, Air Intake, and Fuel System as in the car. The simulation should be so close that you have high confidence that the engine will run the same in the car as on the dyno. Wouldn't it be best to sort out application problems on the dyno using an effective simulation of the vehicle? Sort out fuel problems by running the exact same fuel system taken out of the car into the dyno cell including the fuel cell.

Our New Dyno Standard specifies that a dyno test cell be set up like a carefully controlled lab experiment, which provides a close simulation of the actual application. Now, dyno testing would be Really Meaningful.

There are a lot of ways of doing things wrong and only a few ways of doing things right.

BACK to INDEX

Discussion on INERTIA Corrections in DYNO Testing: Important Updates here, Oct00

Dyno testing of Engines has changed very little over the last 50 years (Really). The way of thinking about dyno testing also has changed very little. Tradition RULES. (Using computers has not changed the Traditional methods and Mentality). In doing the Traditional Static, or steady state, 'Pull' there is no effect of inertia on the Torque gauge. The Traditional Location of the Torque sensor is on the dyno, which is otherwise free to rotate. The total combination of engine, flywheel, drive shaft, and Dyno represents a lot of Moment of Inertia (or just Inertia) that resists changes in RPM. Rotational Inertia is Mass times the Square of the distance from the center of rotation. A heavy Drive shaft has very little Inertia compared to a Flywheel, of the Same weight, but having more Mass far from the center of rotation. The DEPAC System can test engines in a very Untraditional manor involving changing the load and sweeping the Speed of the engine over a desired range. In doing this the system can Statistically Accumulate very accurate engine information and present it in a graphical curve format. This system 'sees' Everything the engine does during the sweep test and can draw a True curve, which is very representative of true engine performance.

In Sweep testing an engine, there can be Large, Erratic errors caused by the Inertia, Absorbing power from the engine on acceleration and releasing it on deceleration. The Faster the sweep, the Larger the errors.The Torque sensor reads Lower on acceleration and Higher on deceleration. Doubling the sweep Rate will double the errors. You always have sweep rate variations and these cause the inertia torque spikes. The Goal is to obtain Sweep testing results for Torque, Free of the errors caused by the Inertia.We should all know that any extra rotational inertia will reduce the power delivered to the drive wheels on rapid acceleration. Making the engine/driveline components as light as possible, and still finish the race, is a worthy goal. But we Must accept a certain minimum inertia for Durability and acceptable Reliability.

In general, Infurnal combustion engines produce different Torque curves when sweeping Up and Down at different rates and then act differently at Steady State. A Turbo-Charged engine has a very obvious Turbo Lag and can severly affect applications that require a rapid acceleration. More subtle are the effects of liquid fuel 'streams' in engines using a Intake distribution manifold, that affect performance sweeping up and down. (There are several SAE Papers concerning these effects).

This is why we recommend testing engines to simulate the application. Drag racing engines Need to be Swept-up at a rate that you will experience at the strip. At these fast rates, the uncorrected Inertia errors Will mask the real changes you need to see. Dyno systems with 'Speed Control' really aren't much better than manual control (even when perfectly tuned), and when mis-adjusted, can be Worst. (Our new Standards Encourage Smooth Load Changes, NOT Speed Control). When the DEPAC system was first introduced back in 1986 we recommended using a Simple Inertia factor as a means to make the manual dyno more 'Operator Proof' and to help fix the erratic control of 'computer controlled' dynos. These latter controlled dynos are not as smooth as the manufacturers suggest. We prefer an Open loop, turn the valve smoothly, method that changes the Torque Load in a smooth, predictable way.

It is a Basic Law of Motion that Inertia Times Sweep Rate Equals a predictable Torque error on the Traditional Dyno torque sensor. You need a system that can accurately sense the sweep rate to get the most accurate inertia correction. Its then great to sweep at different rates and still See small changes that would otherwise be covered up by uncorrected inertia errors. There are other methods for measuring the engine's True Torque without having any inertia errors. (see below)

 Manually setting a Fixed 'Inertia Factor', properly, did a good job of making results more consistent. But it was very difficult for most to understand how to set, or to 'see' the errors, if not set properly (after all, we never had to set this when doing steady state pulls). Proper setting of this important factor is not easy, even when you have a good handle on the problem. Unstable running engines can make it impossible to compare the changes and find an optimum setting from a simple Sweep Up, then Down. Now, We are providing a much more advanced system for Setting a TRUE Inertia Correction. With a properly set inertia factor, you will have the True engine performance curves. True Curves that show real engine changes with Any sweep rate (even Static) and True Curves that Accurately show a badly running engine due to fuel, ignition, or other problems. If a Curve is Smooth, its because the engine is Running Smoothly! Non-smooth, Non-repeatable sweep curves show an engine in trouble and needing work.

The Method used to detect and set this factor requires Very Accurate information on Sweep rate and measured torque. This is where the superior accuracy of the DEPAC System shines. Of course, we encourage measuring Engine Torque, Properly, off the Engine block, where there are NO Inertia errors to correct. (Well,,, changing a Standard takes time,, even generations). This new Technology is a significant breakthrough.

Removing Inertia Errors is Very Significant, for you can then clearly SEE what the engine is doing at different sweep rates, Up and Down, compared to a slow, near, static sweep. Step tests do Not show the engine's behavior under acceleration. You can easily see problems that would be totally masked by Inertia errors. But, Please don't confuse the results as being true to represent power delivered to a load with Inertia being applied. For this you can simply tell the DELINK Program that you have zero correction to see the real performance delivered to the load with inertia present. (since, inertia corrections can be changed after the test, now). You can allow the system to analyze and suggest an optimum Inertia setting but any setting can be Changed by the user. You could look at your runs set to a zero factor and see what the errors would be if they were not corrected. A good demonstration of why we need and use inertia corrections.

DEPAC system can accurately measure Sweep rates to 3 digits (.1%), so we can 'see' a difference between 634 and 635 RPM/Sec. Having accurate information on sweep rate is important to get the most accurate inertia correction within our LINK vers 3 program.

 So, The Updated DEPAC System can provide a near Automatic Inertia correction setting with high precision, with a special Program chip and the new DELINK PC program Version 3.0. The previous method could only change the Inertia factor (in the blue box) in steps like 11, 12, 13. The new system can set a proper variable correction that's much finer, like 12.671, and can remember these settings in the program's INF files for the different types of engine combinations you test. We now do rapid load changes arround a static RPM point and use the playback of the Torque Machine-Gun dots to see if inertia is properly corrected. If not you will see these MGUN dots tracing out a loop. When correct these dots just plot up and down on top of each other on the torque curve.


 

CONCERNING Atmospheric CORRECTION FACTORS:

There are Many FACTORs that Affect the Performance of an Internal Combustion Engine.

CONCERNED here are the AIR INLET CONDITIONS, Correction Factors and Procedures. In the Standard Correction Factor Formula there are two MAIN FACTORS: Absolute Dry AIR PRESSUREand Air TEMPERATURE.

This is complicated by the Amount of Water Vapor in the Atmosphere (expressed as Partial Vapor Pressure). There is a common reference to Relative Humidity (RH), but this applies primarily to Personal Comfort and NOT to engine testing. The (%) RH can be Very deceiving since at low temperatures a 60% RH has very little water vapor pressure compared to 60% RH at high temperatures. One dyno maker has a table of correction factors based on the (%) RH at 100 DegF. Now if one uses this table with a RH taken at 60 DegF, they will have an Over Correction of 3 to 4%. Maybe that's why using this table is so popular! Absolute Humidity is the Factor important to engine testing.It can be expressed as partial vapor pressure. (The Total Atmospheric Pressure is the Sum of the Partial pressures of all its component gases). Water vapor displaces burnable air (of which most is Nitrogen and only 1/5th is Oxygen), so 5% Water vapor pressure leaves only (100% - 5% = ) 95% Combustible air.

The common Correction Factor standards only consider that water vapor ONLY displaces, or reduces, the burnable Air. What they do not even consider is the AFFECT that Water vapor has on Fuel Vaporization and Cylinder Combustion. A very Hot and Very Humid day (much Water vapor) can affect an engine in ways that are NOT covered by the correction factor (CF) formulas.

Consider the Dilemma of the wide variations of the Factors that are used in the CF formula. You can have HIGH Air Pressure and Temperature and still have the SAME CF as with a LOW Air Pressure and Temperature. Does the engine run the same under these widely different conditions, even though the CF says they should??

Another Dilemmais that the Mathematical Exponent applied to the ratio of the Absolute Inlet air temperature to the Standard was determined by Experiment! The Square Root (or exponent 0.5) was simply CHOSEN as a convenient COMPROMISE as experimental results had indicated different Exponents depending on the types of engines being tested, the accuracy of the dynos, and the different test conditions. In Conclusion, it is almost invalid to use ANY Correction Factor at all, except for ONLY small changes from the Standard Conditions. But many use the standard CF formula in ANY Condition as if it were Absolutely Perfect.

DEPAC Dyno Systems RECOMMENDS that VALID engine testing should be conducted under CONTROLLED Air Conditions (which, preferably, are similar to the expected engine operating conditions). COMPARISONS should ONLY be made between tests conducted under the SAME Conditions. Notice that we are NOT talking about using a correction Factor.

The Most important Factor to Control is AIR TEMPERATURE, since its effect is the LEAST Understood and has a DIRECT affect on Air/Fuel Vaporization and Cylinder Combustion Dynamics. Test conditions NEED to be carefully documented, so that other testing can be made under the SAME Conditions for a TRUE Comparison of Results. Air Pressure Can be Allowed to vary over a limited range and a correction can be applied for this, which is a simple DIRECT Ratio, But the Air Temperature MUST be Closely Controlled.The Partial pressure of the water vapor can't be controlled as easily, SO tests should be made to find its effect on different engine types.

Testing Engines in widely different Atmospheric conditions and then Applying a MagicalCorrection Factor to predict the performance to a Standard Condition has been THE accepted procedure. This is easy to do but isn't very meaningful. Controlling the Test Conditions is the proper procedure for any meaningful engine testing.

In the U.S. there are two Main Reference Atmospheric Standards:The Oldest is SAE J607which has a Clear Origin from Aviation. The Barometer standard is 29.92" mercury, which is the Standard Mean Sea-Level pressure, and the Air temperature of 60 DegF, a standard for Aviation Air Density.

In the 80's the SAE realized this Old standard was not meaningful and changed the conditions to a more Realistic 29.53" Mercury of 1% Humid Air at 77 Deg F (25C), (J1349 and J1985). Racers still prefer J607 since it Gives them a 4% higher Corrected Number. We prefer the More Realistic J1985 Standard. (unless you are setting up a Snowmobile to run the Finals at Thief River Falls, this Winter)

BACK to INDEX